The ontology design for solving computational plasma physics problems on supercomputers

Computational plasma physics problems are a wide field with its own hardware, software and scheduling strategies. There is a set of physical phenonena, mathematical equations, numerical approaches, programming strategies and architecture concepts directly followed by each other. This means that an efficient code for the plasma physics must involve all the subjects, which require a clear representation of the relationships amongst physical, mathematical and computer science concepts — which is, the ontology. In this paper it is shown how physics determines equations being used, how the equations define numerical methods, and how methods enable programming strategies to form an architecture-efficient implementation.